Digital image analysis of plus disease in retinopathy of prematurity

Retinal Vessel Local Tortuosity under a Macula-to-Optic Disc Central-Framing Change

Some ocular and cardiovascular diseases can be detected through the increased tortuosity of retinal blood vessels. Objective tortuosity measures can be obtained from digital image analysis of a retinography. This study tested a set of local tortuosity indices under a change in the frame center (macula, optic disc) of the eye fundus image. We illustrate the effects of such a change on 40 pairs of vessels evaluated with eight tortuosity indices. We show that the frame center change caused significant differences in the mean values of the vast majority of the tortuosity indices analyzed. The index defined as the ratio of the curvature to the arc length of a vessel segment proved to be the most robust in relation to a frame center change. Experimental results obtained from the analysis of clinical images are provided and discussed.

Automatic Grading of Retinal Blood Vessel in Deep Retinal Image Diagnosis

Automatic grading of retinal blood vessels from fundus image can be a useful tool for diagnosis, planning and treatment of eye. Automatic diagnosis of retinal images for early detection of glaucoma, stroke, and blindness is emerging in intelligent health care system. The method primarily depends on various abnormal signs, such as area of hard exudates, area of blood vessels, bifurcation points, texture, and entropies. The development of an automated screening system based on vessel width, tortuosity, and vessel branching are also used for grading. However, the automated method that directly can come to a decision by taking the fundus images got less attention. Detecting eye problems based on the tortuosity of the vessel from fundus images is a complicated task for opthalmologists. So automated grading algorithm using deep learning can be most valuable for grading retinal health. The aim of this work is to develop an automatic computer aided diagnosis system to solve the problem. This work approaches to achieve an automatic grading method that is opted using Convolutional Neural Network (CNN) model. In this work we have studied the state-of-the-art machine learning algorithms and proposed an attention network which can grade retinal images. The proposed method is validated on a public dataset EIARG1, which is only publicly available dataset for such task as per our knowledge.

Automated identification of retinopathy of prematurity by image-based deep learning

BACKGROUND

Retinopathy of prematurity (ROP) is a leading cause of childhood blindness worldwide but can be a treatable retinal disease with appropriate and timely diagnosis. This study was performed to develop a robust intelligent system based on deep learning to automatically classify the severity of ROP from fundus images and detect the stage of ROP and presence of plus disease to enable automated diagnosis and further treatment.

METHODS

A total of 36,231 fundus images were labeled by 13 licensed retinal experts. A 101-layer convolutional neural network (ResNet) and a faster region-based convolutional neural network (Faster-RCNN) were trained for image classification and identification. We applied a 10-fold cross-validation method to train and optimize our algorithms. The accuracy, sensitivity, and specificity were assessed in a four-degree classification task to evaluate the performance of the intelligent system. The performance of the system was compared with results obtained by two retinal experts. Moreover, the system was designed to detect the stage of ROP and presence of plus disease as well as to highlight lesion regions based on an object detection network using Faster-RCNN.

RESULTS

The system achieved an accuracy of 0.903 for the ROP severity classification. Specifically, the accuracies in discriminating normal, mild, semi-urgent, and urgent were 0.883, 0.900, 0.957, and 0.870, respectively; the corresponding accuracies of the two experts were 0.902 and 0.898. Furthermore, our model achieved an accuracy of 0.957 for detecting the stage of ROP and 0.896 for detecting plus disease; the accuracies in discriminating stage I to stage V were 0.876, 0.942, 0.968, 0.998 and 0.999, respectively.

CONCLUSIONS

Our system was able to detect ROP and differentiate four-level classification fundus images with high accuracy and specificity. The performance of the system was comparable to or better than that of human experts, demonstrating that this system could be used to support clinical decisions.

A computer-aided diagnosis system for plus disease in retinopathy of prematurity with structure adaptive segmentation and vessel based features

Retinopathy of Prematurity (ROP) is a blinding disease affecting the retina of low birth-weight preterm infants. Accurate diagnosis of ROP is essential to identify treatment-requiring ROP, which would help to prevent childhood blindness. Plus disease, which characterizes abnormal twisting, widening and branching of the blood vessels, is a significant symptom of treatment requiring ROP. In this paper, we have developed and evaluated a computer-based analysis system for objective assessment of plus disease in ROP, which best mimics the clinical method of disease diagnosis by identifying unique vessel based features. The proposed system consists of an initial segmentation stage, which will efficiently extract blood vessels of varying width and length by utilizing structure adaptive filtering, connectivity analysis and image fusion. The paper proposes the usage of additional retinal features namely leaf node count and vessel density, to portray the abnormal growth and branching of the blood vessels and to complement the commonly used features namely tortuosity and width. The test results show a better classification of plus disease in terms of sensitivity (95%) and specificity (93%), emphasizing the superiority of the proposed segmentation algorithm and vessel-based features. An additional advantage of the proposed system is that the process of selection of relevant vessels for feature extraction is fully automated, which makes the system highly useful to the non-physician graders, owing to the unavailability of a sufficient number of ROP specialists.

Quantitative Analysis of Conjunctival and Retinal Vessels in Fabry Disease

Fabry Disease (FD) is a rare X-linked lysosomal storage disorder characterized by systemic and ocular involvement. It has been described an increasing in retinal and conjunctival vessel tortuosity and this feature represents an important marker for the disease. Currently, there is not an objective method to measure and quantify this parameter. We tested a new semi-automatic software measuring retinal and conjunctival vessel tortuosity from eye fundus and conjunctival digital images in a group of FD patients. We performed an observational case-control study evaluating three mathematical parameters describing tortuosity (sum of angle metric [SOAM], product of angle distance [PAD], triangular index [I2e]) obtained from fundus and conjunctival pictures of 11 FD patients and 11 age and sex-matched controls. Both eyes were considered. Mann-Whitney test was used to compare the FD group versus the control group and, within the FD group, male versus female patients. Linear regression analysis was performed to evaluate the possible association of retinal and conjunctival vessels tortuosity parameters with age and with specific markers of systemic disease's progression. The tortuosity parameters (SOAM, PAD and I2e) were significantly higher in retinal vessels and in conjunctival nasal vessels in FD patients in comparison with the controls (p=0.003, p=0.002, p=0.001 respectively for retina) (p=0.023, p=0.014, p=0.001 respectively for nasal conjunctiva). No significant association was found between retinal and conjunctival tortuosity parameters and increasing age or systemic involvement markers. Vessel tortuosity represents an important clinical manifestation in FD. A computer-assisted analysis of retinal and conjunctival vasculature demonstrated an increased vessels tortuosity in patients affected by Fabry disease. This non-invasive technique might be useful to help the diagnosis in early stages, to establish disease severity and monitor its progression.

Automated Analysis for Retinopathy of Prematurity by Deep Neural Networks

Retinopathy of Prematurity (ROP) is a retinal vasproliferative disorder disease principally observed in infants born prematurely with low birth weight. ROP is an important cause of childhood blindness. Although automatic or semi-automatic diagnosis of ROP has been conducted, most previous studies have focused on "plus" disease, which is indicated by abnormalities of retinal vasculature. Few studies have reported methods for identifying the "stage" of the ROP disease. Deep neural networks have achieved impressive results in many computer vision and medical image analysis problems, raising expectations that it might be a promising tool in the automatic diagnosis of ROP. In this paper, convolutional neural networks with a novel architecture are proposed to recognize the existence and severity of ROP disease per-examination. The severity of ROP is divided into mild and severe cases according to the disease progression. The proposed architecture consists of two sub-networks connected by a feature aggregate operator. The first sub-network is designed to extract high-level features from images of the fundus. These features from different images in an examination are fused by the aggregate operator, then used as the input for the second sub-network to predict its class. A large data set imaged by RetCam 3 is used to train and evaluate the model. The high classification accuracy in the experiment demonstrates the effectiveness of the proposed architecture for recognizing the ROP disease.

COMPUTER ASSISTED RETINAL VESSEL TORTUOSITY EVALUATION IN NOVEL MUTATION FABRY DISEASE: Towards New Prognostic Markers

PURPOSE

Fabry disease is a rare lysosomal storage disorder with systemic involvement. The authors report on a large Fabry family with GLA p.M187R mutation and exhaustive ophthalmologic assessment.

METHODS

Comprehensive systemic evaluation and genetic diagnosis were performed. Ophthalmologic evaluation included intraocular pressure/visual acuity measurement, refractometry, slit lamp examination, retinography, and optical coherence tomography. Three parameters quantified retinal vessel tortuosity: sum of angle metrics, product of angle distance, and triangular index. Calculations were semiautomatized using dedicated software.

RESULTS

Ten individuals (2 males and 8 females) were described. Seventy-five percent had retinal vessel tortuosity. One hundred percent had cornea verticillata. Perimacular vessels were predominantly involved. The correlation between the right and left eye tortuosity measurements was very tight. A significant correlation between retinal vessel tortuosity and systemic severity measured by general Mainz Severity Score Index (MSSI), renal MSSI, and neurological MSSI but no cardiac MSSI was observed. Right sum of angle metrics value was an independent statistical predictor of the general-MSSI score in presence of age.

CONCLUSION

p.M187R mutation causes a severe systemic and ophthalmologic phenotype, in both male and female patients. Semiautomatic assessment of retinal vessel tortuosity is an objective and reproducible tool. All three parameters of tortuosity are closely associated with Fabry severity scores. Studies of larger series are being awaited to establish the role of retinal vessel tortuosity as a noninvasive marker of disease progression.

The Novel Evidenced Assessment of Tortuosity system: interobserver reliability and agreement with clinical assessment

PURPOSE

Computer-assisted assessment of vessel tortuosity is clinically useful in retinopathy of prematurity (ROP). However, poor image quality is often prohibitive for accurate segmentation by fully automated systems and semi-automated systems are prone to unreliability. In the present work, we describe a method of retinal vessel tortuosity measurement by means of purpose-built image analysis software that does not require high image quality yet is also reliable.

METHODS

Images were obtained from neonates at risk of ROP with Retcam Shuttle(®) . Individual vessels were assessed with the semi-automated Novel Evidenced Assessment of Tortuosity (NEAT) system by two masked experimenters. Scores were compared to assess reliability. They were also compared against clinical scoring of individual vessels by two ROP screeners to assess relationship with clinical assessment. In a second image cohort, the mean of the most tortuous vessel in each of four quadrants in each eye (NEAT-O) was compared against the documented gold standard clinical grading of plus disease.

RESULTS

Reliability of the NEAT system for 50 individual vessels using Bland-Altman plots was excellent. NEAT tortuosity scores for 50 individual vessels compared to clinical scoring showed strong correlation (0.706). Correlation between the NEAT-O score for average tortuosity and gold standard for 167 eyes was modest (0.578).

CONCLUSIONS

The NEAT system is intuitive, user-friendly and robust enough to be clinically useful in poor-quality images. It allows for a rapid, valid and reliable assessment of tortuosity of individual vessels and produces a tortuosity score that correlates well with severity of plus disease.

Enhancing Image Characteristics of Retinal Images of Aggressive Posterior Retinopathy of Prematurity Using a Novel Software, (RetiView)

Purpose. To report pilot data from a novel image analysis software "RetiView," to highlight clinically relevant information in RetCam images of infants with aggressive posterior retinopathy of prematurity (APROP). Methods. Twenty-three imaging sessions of consecutive infants of Asian Indian origin with clinically diagnosed APROP underwent three protocols (Grey Enhanced (GE), Color Enhanced (CE), and "Vesselness Measure" (VNM)) of the software. The postprocessed images were compared to baseline data from the archived unprocessed images and clinical exam by the retinopathy of prematurity (ROP) specialist for anterior extent of the vessels, capillary nonperfusion zones (CNP), loops, hemorrhages, and flat neovascularization. Results. There was better visualization of tortuous loops in the GE protocol (56.5%); "bald" zones within the CNP zones (26.1%), hemorrhages (13%), and edge of the disease (34.8%) in the CE images; neovascularization on both GE and CE protocols (13% each); clinically relevant information in cases with poor pupillary dilatation (8.7%); anterior extent of vessels on the VNM protocol (13%) effecting a "reclassification" from zone 1 to zone 2 posterior. Conclusions. RetiView is a noninvasive and inexpensive method of customized image enhancement to detect clinically difficult characteristics in a subset of APROP images with a potential to influence treatment planning.

Ophthalmic experience over 10 years in an observational nationwide Danish cohort of Fabry patients with access to enzyme replacement

PURPOSE

Enzyme replacement therapy (ERT) was offered from year 2001 to patients with Fabry disease. The ophthalmic experience was analysed, as part of a general 10-year status.

METHODS

A retrospective observational series comprising 39 patients (25 females, 14 males) closely followed by the endocrinologists, and with regular ophthalmic control. Time of inclusion was when the option of ERT was started, at age 11-60 years. Eye data (standard eye examination, including retinal imaging) were incomplete in five, due to death or non-attendance, and five patients had refused treatment.

RESULTS

Vision was normal throughout, except in two young males with total unilateral central retinal artery occlusion, prior to and during enzyme replacement, respectively. Cornea verticillata and conjunctival vessel ectasies were common. Tortuosity of retinal arterioles and venules was recorded in eight and 18 patients, respectively, and phlebopathy in 22, although generally without evidence of loss of vessel wall integrity. Systemic vascular lesions with or without loss of function were recorded for kidney (n = 23), heart (n = 17) and brain (n = 7), and an association was suggested between nephropathy and abnormal morphology of retinal vessels.

CONCLUSIONS

Thirteen of 32 patients on ERT showed a reduction of corneal deposits over the study period. Abnormal ocular vessel morphology was a frequent finding. In contrast to the function loss related to systemic ischaemic lesions, we found no indication of impairment of visual parameters in 37. Compared to other Fabry series, two of 39 patients with serious unilateral occlusive retinal disease may appear a high number. The presence of retinal tortuosity is discussed, possibly reflecting haemodynamic events related to vessel wall deposits, but could also be 'constitutional', as part of the Fabry inheritance.

Intravitreal low-dosage bevacizumab for retinopathy of prematurity

PURPOSE

To report on the therapeutic effect of intravitreal low-dose bevacizumab for treatment for retinopathy of prematurity (ROP).

METHODS

The single-centre retrospective, non-comparative case series study included all infants who consecutively underwent intravitreal injection of 0.375 mg bevacizumab (0.03 ml) under light sedation in topical anaesthesia as therapy of ROP in zone I or zone II.

RESULTS

The clinical charts of 29 patients (57 eyes) with a median birth weight of 630 g (range: 290-1390 g) and median gestational age of 25 + 1 weeks (range: 23 + 1-30 weeks) were reviewed. Six children (12 eyes) were graded as ROP with zone I retinopathy and plus disease. The 23 remaining infants had extraretinal neovascularizations in zone II or partly zone I. The intravitreal bevacizumab injection was injected at a median age of 12 + 1 weeks (range: 7 + 4-21 + 4), the median follow-up was 4.2 months (range: from 3 days to 45.1 months). In all eyes treated, a regression of plus disease occurred within two to six days, retinal neovascularizations regressed within 2-3 weeks and pupillary rigidity improved. None except one child in exceptionally bad general health conditions needed a second intravitreal bevacizumab injection. In none of the infants, any ophthalmologic side-effects of the bevacizumab application were detected during the follow-up period.

CONCLUSIONS

The intravitreal injection of a low dose of 0.375 mg bevacizumab showed a high efficacy as treatment for ROP. The question arises whether the low dosage of bevacizumab as compared to the dosage of 0.625 mg bevacizumab may be preferred.

Fluorescein angiography and retinal vascular development in premature infants

OBJECTIVE

To investigate the role of fluorescein angiography (FA) in the management of retinopathy of prematurity (ROP) in preterm newborns.

METHODS

An observational case series of 13 extremely low birth weight infants. From September 2009 to March 2010, 13 newborn infants with a gestational age <29 weeks end/or birth weight <1000 g underwent serial fluorescein angiography with RetCam (Clarity, Pleasanton, CA) every 2 weeks. The fluorescein angiograms were examined to optimize the timing of diagnosis of ROP and to investigate development of retinal and choroidal vascularization.

RESULTS

There were no side effects related to FA. Variable features of retinal and choroidal circulation in preterm infants with a high risk of developing ROP were noted. FA allows vessels branching at the junction between vascular and avascular retina (V-Av junction) to be viewed easily and shows the ROP findings that sometimes cannot be seen by indirect ophthalmoscopy. Dye leakage is the most significant sign of progression to severe ROP or the need for surgery in newborn babies with ROP.

CONCLUSIONS

RetCam-assisted intravenous FA is safe and allows a more objective assessment of the ROP stage and zone.

Computer assisted evaluation of retinal vessels tortuosity in Fabry disease

PURPOSE

Fabry Disease (FD) is a rare X-linked metabolic disorder characterized by diffuse deposition of sphingolipids in many tissues. Retinal vessel tortuosity is a common ocular manifestation in FD and may represent a useful marker for the disease. Unfortunately its clinical evaluation is poorly reproducibile and alternative means of evaluation may be of interest. We tested a new semi-automatic software measuring retinal vessel tortuosity from eye fundus digital images in a group of FD patients.

METHODS

Observational case-control study evaluating four mathematical parameters describing tortuosity (relative length, sum of angle metric [SOAM], product of angle distance [PAD], triangular index) obtained from fundus pictures of 35 FD patients and 35 age-matched controls. Only the right eye was considered in order to reduce bias. Mann-Whitney test was used to compare the FD group versus the control group, males versus females and patients with versus without clinically identified retinal vessels tortuosity in the FD group. Linear regression analysis was performed on a subgroup of patients to evaluate the possible association of retinal vessels tortuosity parameters with age and with markers of systemic disease's progression.

RESULTS

Three parameters (SOAM, PAD and triangular index) were significantly higher in FD patients in comparison with the controls (p < 0.0001, p = 0.001, p = 0.002 respectively). In the FD group the same three parameters showed higher values in hemizygous males than in heterozygous females ((p < 0.0001, p = 0.002, p < 0.0001 respectively).

CONCLUSION

A computer assisted analysis of retinal vasculature demonstrated an increased vessels tortuosity in FD patients. The technique might be useful to establish disease severity and monitor its progression.

Retinal vessel tortuosity measures and their applications

Structural retinal vascular characteristics, such as vessel calibers, tortuosity and bifurcation angles are increasingly quantified in an objective manner, slowly replacing subjective qualitative disease classification schemes. This paper provides an overview of the current methodologies and calculations used to compute retinal vessel tortuosity. We set out the different parameter calculations and provide an insight into the clinical applications, while critically reviewing its pitfalls and shortcomings.